A method of block encoding of an image includes partitioning an image into a plurality of blocks each having a plurality of pixels corresponding to a first matrix, obtaining a first block from the plurality of blocks, rearranging pixels of the first block for generating a second block according to an arrangement rule, performing wavelet transformation for the second block for generating a third block having a plurality of wavelet coefficients corresponding to the first matrix, rearranging the wavelet coefficients of the third block for generating a fourth block according to the arrangement rule, quantizing the fourth block, and encoding a quantized result of the fourth block.
Legal claims defining the scope of protection, as filed with the USPTO.
1. A method of block encoding of an image comprising: partitioning an image into a plurality of blocks each having a plurality of pixels corresponding to a first matrix; obtaining a first block from the plurality of blocks; rearranging pixels of the first block for generating a second block according to an arrangement rule; performing wavelet transformation for the second block for generating a third block having a plurality of wavelet coefficients corresponding to the first matrix; rearranging the wavelet coefficients of the third block for generating a fourth block according to the arrangement rule; quantizing the fourth block; and encoding a quantized result of the fourth block.
2. The method of claim 1 , wherein the first matrix is a 4×4 matrix.
3. The method of claim 1 , wherein the arrangement rule is mapping a block to a second matrix and replacing a value of the block corresponding to a first coordinate of the second matrix with a value of the block corresponding to a second coordinate of the second matrix.
4. The method of claim 3 , wherein the second matrix is a 4×4 matrix.
5. The method of claim 4 , wherein the first coordinate is ((2i+m), (2j +n)) and the second coordinate is ((2m+i), (2n+j)), where m, n, i, j are 0 or 1.
6. The method of claim 1 , wherein performing wavelet transformation for the second block for generating the third block is partitioning the second block into a plurality of equal-size sub-blocks and performing Haar wavelet transformation for the plurality of equal-size sub-blocks for generating the third block.
7. The method of claim 6 , wherein each of the plurality of equal-size sub-blocks comprises pixels corresponding to a third matrix.
8. The method of claim 7 , wherein the third matrix is a 2×2 matrix.
9. The method of claim 6 , wherein performing Haar wavelet transformation for the plurality of equal-size sub-blocks is applying a low-pass-filtering matrix and a high-pass-filtering matrix for obtaining a low-pass wavelet coefficient and a plurality of high-pass wavelet coefficients of each sub-block.
10. The method of claim 9 , wherein the high-pass-filtering matrix is [−1 1] and the low-pass-filtering matrix is [1 1].
11. The method of claim 1 , wherein the fourth block comprises a high-correlation sub-block and a plurality of low-correlation sub-blocks, wherein the high-correlation sub-block comprises a plurality of low-pass wavelet coefficients of the third block and the plurality of low-correlation sub-blocks comprise a plurality of high-pass wavelet coefficients of the third block.
12. The method of claim 11 , wherein the high-correlation sub-block and the plurality of low-correlation sub-blocks are corresponding to a 2×2 matrix.
13. The method of claim 1 , wherein quantizing the fourth block comprises: determining a refinement scale of the fourth block; and quantizing the fourth block according to the refinement scale of the fourth block.
14. The method of claim 13 , wherein determining the refinement scale of the fourth block is mapping the fourth block to a fourth matrix and determining the refinement scale of the fourth block according to a plurality of inequalities.
15. The method of claim 14 , wherein the fourth matrix is a 4×4 matrix and the plurality of inequalities comprise: (abs(p(1,0)−abs(p(0,0))<127), (abs(p(0,1)−abs(p(0,0))<127), (abs(p(1,1)−abs(p(1,0))<127), (abs(p(1,1)−abs(p(0,1))<127), (abs(p(0,2))<127), (abs(p(0,3))<127), (abs(p(1,2))<127), (abs(p(0,3)−abs(p(0,2))<127), (abs(p(1,3)−abs(p(1,2))<127), (abs(p(1,2)−abs(p(0,2))<127), (abs(p(1,3)−abs(p(0,3))<127), (abs(p(2,0))<127), (abs(p(2,1))<127), (abs(p(3,0))<127), (abs(p(2,1)−abs(p(2,0))<127), (abs(p(3,1)−abs(p(3,0))<127), (abs(p(3,0)−abs(p(2,0))<127), (abs(p(3,1)−abs(p(2,1))<127), (abs(p(2,2))<127), (abs(p(2,3))<127), (abs(p(3,2))<127), (abs(p(2,3)−abs(p(2,2))<127), (abs(p(3,3)−abs(p(3,2))<127), (abs(p(3,2)−abs(p(2,2))<127), and (abs(p(3,3)−abs(p(2,3))<127), where abs(x) represents an absolute value of x, and p(a, b) represents a pixel value of the fourth block corresponding to a coordinate (a, b) of the matrix.
16. The method of claim 13 , wherein quantizing the fourth block according to the refinement scale of the fourth block comprises: utilizing a first bit allocation scheme for quantizing the fourth block when the refinement scale of the fourth block is higher than a predetermined scale; and utilizing a second bit allocation scheme for quantizing the fourth block when the refinement scale of the fourth block is lower than the predetermined scale.
17. The method of claim 16 , wherein the total number of bits of the first bit allocation scheme is the same as the total number of bits of the second bit allocation scheme, and the number of bits of the first bit allocation scheme corresponding to a low-frequency part of the fourth block is larger than the number of bits of the second bit allocation scheme corresponding to the low-frequency part of the fourth block.
18. The method of claim 16 , wherein quantizing the fourth block according to the refinement scale of the fourth block further comprises adjusting the first bit allocation scheme and the second bit allocation scheme according to a compression ratio.
19. The method of claim 1 , wherein encoding a quantized result of the fourth block comprises performing round-off operation, determining a prediction direction, and performing differential pulse code modulation for the quantized result of the fourth block.
20. A method of block decoding of an image comprising: receiving bits of a first block; determining a bit allocation of the first block and an encoding prediction direction; decoding the first block into a second block having a plurality of wavelet coefficients corresponding to a first matrix according to the encoding prediction direction; rearranging the plurality of wavelet coefficients of the second block for generating a third block according to an arrangement rule; performing inverse wavelet transformation for generating a fourth block having a plurality of pixels corresponding to the first matrix; rearranging the plurality of pixels of the fourth block for generating a fifth block according to the arrangement rule; and outputting the fifth block.
21. The method of claim 20 , wherein the first matrix is a 4×4 matrix.
22. The method of claim 20 , wherein the arrangement rule is mapping a block to a second matrix and replacing a value of the block corresponding to a first coordinate of the second matrix with a value of the block corresponding to a second coordinate of the second matrix.
23. The method of claim 22 , wherein the second matrix is a 4×4 matrix.
24. The method of claim 23 , wherein the first coordinate is ((2i+m), (2j+n)) and the second coordinate is ((2m+i), (2n+j)), where m, n, i, j are 0 or 1.
25. The method of claim 20 , wherein performing inverse wavelet transformation for the third block for generating the fourth block is partitioning the third block into a plurality of equal-size sub-blocks and performing inverse Haar wavelet transformation for the plurality of equal-size sub-blocks for generating the fourth block.
26. The method of claim 25 , wherein each of the plurality of equal-size sub-blocks comprises pixels corresponding to a third matrix.
27. The method of claim 26 , wherein the third matrix is a 2×2 matrix.
Cooperative Patent Classification codes for this invention. Click any code to explore related patents in that topic.
August 31, 2006
April 20, 2010
Browse 5M+ US patents with plain-English claim translations and AI-generated analysis.